Vacuum Tubes#

The vacuum tube was the first device to switch electronically — with no moving parts. Where a relay throws a metal armature to make or break a circuit, a tube controls a stream of electrons flying through a vacuum. Nothing mechanical moves, so switching happens at the speed of electron flow rather than the speed of a spring. That single change took switching from milliseconds to microseconds, and it is what made genuinely fast computation possible.

How a Vacuum Tube Works#

A tube works by thermionic emission. A heated cathode boils electrons off its surface; a positively charged plate (anode) across an evacuated envelope attracts them, and current flows. With just those two electrodes the device is a diode — it conducts in one direction only, which is why the earliest tubes were rectifiers.

The breakthrough was adding a third electrode. In a triode, a wire mesh control grid sits between cathode and plate, and a small voltage on the grid throttles the electron stream — a slightly negative grid repels electrons and reduces plate current; a more negative grid cuts it off entirely. Because a tiny grid voltage controls a large plate current, the triode amplifies. Later tubes (tetrodes, pentodes) added further grids to improve gain and reduce unwanted capacitance, but the principle never changed: a small input voltage controls a much larger output current. That is transconductance, and it is the same idea that carries directly into the transistor, so the mental model transfers rather than being discarded.

Vacuum Tubes as Logic#

For logic, the tube is not used as a linear amplifier but slammed between two extremes: the grid is driven below cutoff (plate current off) or hard into conduction (plate current on). Those two states are the 1 and 0, switched electronically in microseconds; the linear region in between is for analog amplifiers, and a gate deliberately avoids resting there.

Combinational gates were built by pairing diodes with tubes: diode networks form AND and OR by steering current through whichever path is enabled, and a triode or pentode stage inverts and restores the signal level so gates can be cascaded without the levels decaying — the same division of labor that later became DTL.

The more consequential invention was memory. In 1918 the Eccles–Jordan circuit cross-coupled two triodes so that each held the other in place: the pair rests in one of two stable states and stays there until a pulse flips it. That bistable circuit is the flip-flop — electronic memory, invented in tubes, decades before the transistor. Sequential logic became electronic here. The first electronic computers counted and stored numbers in banks of these tube flip-flops; ENIAC’s decade counters, for instance, were ring counters of ten flip-flops each. Spotting two cross-coupled active devices that feed back to hold each other is spotting one bit of storage — a topology identical whether the devices are triodes, bipolar transistors, or CMOS.

Why Tubes Replaced Relays — and Why the Transistor Replaced Tubes#

Tubes displaced relays for computing for one reason: speed. A relay switches in milliseconds and wears out mechanically; a tube switches in microseconds with nothing moving to wear, a thousand-fold speedup that turned automatic calculation into something fast enough to be genuinely useful. For about a decade a tube was simply what an electronic switch was, and the first generation of electronic computers was built on them — Colossus (1943–44) for codebreaking, ENIAC (1945, roughly 18,000 tubes), and their successors.

What tubes gave up drove the search for something better. Every tube has a heater that burns power continuously whether or not it is switching, so thousands of them dissipate kilowatts and demand serious cooling — a limit that alone capped how large tube machines could grow. They are bulky and fragile, glass envelopes with filaments that fail like light bulbs, so in a system of thousands the mean time between failures is dominated by burned-out filaments and a single dead tube can halt the whole machine. Their logic levels are plate swings of hundreds of volts, a world away from the sub-5-volt logic that followed, and they need warm-up time and are microphonic — mechanical vibration modulates the electrodes. The transistor, demonstrated in 1947, did the same switching job solid-state: cool, tiny, rugged, and low-voltage. It made the tube obsolete for logic almost as soon as it could be manufactured in quantity — and the mapping was direct, since the triode’s grid, plate, and cathode become the transistor’s gate/base, drain/collector, and source/emitter. The three-terminal “small input controls large output” device is the same concept in both eras, which is why tube-era circuit topologies, and the word gate itself, carried straight across into semiconductors.

Where Vacuum Tubes Stand Today#

The tube’s logic role ended completely with the transistor, but two threads survive. As a device, the vacuum tube still wins where power and frequency are extreme: high-power RF and microwave transmitters, and the magnetron in every microwave oven, are vacuum tubes, because solid-state does not yet fully match certain very-high-power, high-frequency, high-voltage regimes. This is physics, not nostalgia — but it is generating and switching power, not computing. And in audio, guitar and hi-fi amplifiers still use tubes for the soft-clipping character of their overload, which is the tube as a linear amplifier, not a switch.

What actually persisted from the tube era into modern logic is not the tube but the idea it first made electronic: the cross-coupled Eccles–Jordan bistable is functionally the same latch drawn today in CMOS. The storage primitive behind every register and flip-flop predates the transistor entirely; only the switching device underneath it changed.